JP6784003B2 - Methods, devices, computer-readable storage media, and programs for configuring collaborative simulations for integrated systems - Google Patents

Description

The present invention relates to methods and devices for constructing joint simulations for integrated systems.

Virtual development of technical systems is conventional and enables early system analysis and virtual testing, thus saving time and money. In that case, in each technical area (for example, mechanical engineering, electro-optics, etc.), a model for a specific part is developed by a specific simulation tool, and is simulated and analyzed by itself. The impact of other systems is simply considered limited. Since the development of integrated systems generally requires the coordination of all technical areas or all components, this interaction must also be mapped on a virtual plane. Collaborative simulation offers the possibility to combine models that are decentralized and mimicked. In this case, the binding amount is exchanged during the simulation at a predetermined time point after the so-called macro time step. The characteristics of the simulation tool used and the characteristics of the model determine which coupling algorithm can be used. If the simulation tool or mimicking model (eg, FMI) allows iterative computation steps, iterative (implicit) methods (tight binding) may be utilized. However, in most cases, the simulation tools and simulation model reconfiguration required for this are not assisted, or the simulation must be done in real time, so non-repetitive (explicit) methods must be used. .. However, in the case of inner loops (data dependencies), extrapolation of the binding amount is forced over the actual macro time steps during non-repetitive collaborative simulations, which results in inaccuracy.

The resulting error can be kept negligibly small with proper configuration of the joint simulation, but it must be done manually by the user of the joint simulation according to the prior art. This is time consuming and often cannot be done manually by the collaborative simulation user due to the major complexity of the collaborative simulation.

AT509930A2, entitled "Model-Based Methods and Methods for Quantizing the Quality of Collaborative Simulation Results," describes methods for assessing the quality of collaborative simulation integrated systems. Approximate model descriptions, join information and execution sequences as static meta information are used in combination with the join uncertainty introduced and modeled by extrapolation to determine the quality in advance, i.e. before the start of joint simulation. used.

In joint simulation, the modeled subsystems are combined via the model's input / output parameters into an integrated system. The subsystem is, in this case, an algebraic and / or differential equation system. However, each subsystem is solved or simulated by a unique numerical algorithm over macro time steps, independent of the other subsystems. The solution algorithm is usually selected by the user according to the system of equations to be solved in this case. Each simulation of the subsystem incorporates a specific numerical solution algorithm (solution method). Since data dependencies occur between related subsystems in the network / inner loop, the macro time to be solved by the extrapolation method is the appropriate amount of coupling (input parameter) in parallel or sequentially according to the execution sequence. Must be "estimated" over the steps. This extrapolation is necessary to solve the major causal problems and inevitably introduces the necessary errors.

The error introduced by this inevitable coupling, on the one hand, significantly affects the physical response of the distributed system. For example, when forces are used as coupling quantities in such joint simulations, coupling errors act directly on the integrated system response, for example due to anomalous inertial acceleration.

An object of the present invention is to create and improve the configuration of a joint simulation.

This problem can be solved by a method of constructing a joint simulation for an integrated system according to independent claims and by using an apparatus for constructing a joint simulation for an integrated system.

According to the first aspect of the present invention, a method of constructing a joint simulation (integrated simulation) for an integrated system having at least one first subsystem and a second subsystem. The first partial system has at least one first parameter input and at least one first parameter output, and the first parameter output uses the first solution algorithm based on the first parameter input. The second subsystem has at least one second parameter input and at least one second parameter output. Based on the second parameter input, the second parameter output can be determined using the second solution algorithm.

According to this method, the first subsystem and the second subsystem are coupled by a coupler, and either of the first and second parameter outputs is for the corresponding first and second parameter inputs. A combined network is defined to determine whether it is defined as the amount of binding.

Which first and second parameter output of the corresponding first and second subsystems is the corresponding first and second parameter outputs of the other subsystem (eg, of the other first and second subsystems). Whether it is defined as the amount of coupling for parameter input is determined by the user of the collaborative simulation (sometimes with automatic assistance, eg, "name mapping"). The resulting coupling network determination consists of, for example, all subsystems (so-called vertices), their parameter inputs and parameter outputs, and fixed couplings (edges) of parameter inputs and outputs in the real world inside the joint simulation network. Represents the acquisition of physically related information technology across directed graphs.

Further, the first subsystem information of the first subsystem (for example, direct access, input / output occasional operation, instantaneous frequency, simulation time, etc.) and the second subsystem information of the second subsystem are defined.

For example, there is direct access between the parameter inputs and outputs of the subsystem during mechanical tight binding in the subsystem. For example, the position of the first mass is pre-given by the parameter input, thereby directly accessing the position of the tightly coupled second mass as the parameter output (regardless of the force on this process. is necessary). For example, when the position of the first mass fluctuates, the position of the second mass fluctuates accordingly. This fluctuation does not have to be done exclusively linearly. In general, direct access describes the static association between the parameter input x and the parameter output y of equation y = g (x), where g (x) represents any static map or function. For example, at a predetermined position x of the first mass, a predetermined position y of the second mass can be directly obtained based on a predetermined feature line. The corresponding feature lines may run linearly or, for example, exponentially. Direct access is estimated by the corresponding direct variation of the parameter output with respect to the parameter input.

The input or output ad hoc operation represents the underlying physical regularity of the dynamic process between the parameter input and the parameter output of the subsystem. In mathematics, differential equations are used to describe dynamic processes. For example, when the mass m is moved by the force f as the parameter input and the position fluctuation or the fluctuation of the position x is obtained in the parameter output, the input or the output is operated at any time based on the result of the parameter output related to the parameter input, that is, each other. The physical mathematical relationship to can be inferred. The input / output ad hoc operation is described by the differential equation d ² x / dt ² = F / m in the above example, and characterizes the input or output ad hoc operation between the parameter input and the associated parameter output. If the input / output occasional operation is estimated using the parameter input and the parameter output, it is important to determine the parameters of the underlying differential equation, eg the mass in the above example. In this regard, the corresponding occasional operation or input / output occasional operation can be determined based on the relational expression between the parameter input and the parameter output.

The instantaneous frequency represents the frequency of the coupled signal at a given time point. Different physical effects can result in different instantaneous frequencies in this case. For example, the wheels of an automobile rotate at a constant speed and at a constant rotation speed. The number of wheel revolutions, and thus the instantaneous frequency of this parameter output or parameter input, is reduced during braking. In addition, when a frictional effect occurs, the result is a high frequency portion, which results in a significant variation in instantaneous frequency. If the wheel speed is delayed, the coupling signal has a constant instantaneous frequency. Significant fluctuations in instantaneous frequency can be seen when the wheel speed drops sharply, for example due to the introduction of an anti-block braking system.

Determining partial system information involves obtaining pre-available information, such as combined networks and / or direct access, and information generated during execution time. The actual simulation time of the subsystem is acquired, the input / output ad hoc operation is by the method derived from system identification (eg, through the recursive least squares method, the finite difference method), and the instantaneous frequency is the method derived from signal processing (eg, Hilbert Fang). Estimated by transformation), "direct access" is determined by direct input / output associations (eg, linear or non-linear associations).

In addition, select an execution sequence that determines in which sequence the first and / or second parameter outputs are defined with each other (thus, which first and / or second to solve the causality problem. Determine if the parameter input must be supplemented).

For example, the force represents the first parameter input of the first subsystem, which represents mechanical inertia. The resulting mass position variation is available in the first parameter output, the first parameter output is set as the second parameter input of the second subsystem, and this position variation is In the second subsystem, it becomes the force as the second parameter output (for example, Coulomb force) proportionally, this second parameter output is set again as the first parameter input, and the execution sequence is the joint simulation. It can affect the physical response. If the execution sequence is chosen so that the force must be extrapolated, the extrapolation error is weakened based on the inertia of the mass and acts on the displacement of the mass. On the contrary, if the execution sequence is chosen so that the position of the mass must be extrapolated, the result is a proportionally enhanced error based on the proportional relationship in the second subsystem. It is generated by the parameter output or force of. Execution sequences are therefore important for qualitative, high-value physical integrated system responses.

In addition, in the case of the estimation needed to solve the causality problem, the first and second parameter inputs can be determined using that method during the macro step width (ie, between the time points of coupling). Extrapolation method is determined.

Extrapolation methods affect the physical response of the integrated system in various ways. For example, if the force acting on the mass is extrapolated by 0th order extrapolation (the last known one is kept constant over the actual time step), this will be a staircase curve of force or parameter input .. Depending on the magnitude of the mass, this recoil of force has a significant effect on the system response. No significant disturbance is expected at large masses. Very small masses, on the other hand, are encouraged to move directly, resulting in unwanted vibrations (eg strong vibrations). Create a "smooth" external force curve. Therefore, the choice of extrapolation method is important for qualitative, high-value physical integrated system response.

The corresponding exchange of the first input parameter and the second input parameter and the exchange of the first output parameter and the second output parameter between the first subsystem and the second subsystem are executed at that time. The macro step width is determined by giving the combination time in advance and determining the supplementary range of the first and second input parameters.

The larger the macro step width, the more the parameter inputs must be extrapolated for the future and therefore have a partial significant impact on the physical response of the integrated system. For example, if a force as a parameter input is applied to the mass in the partial system and this parameter input is extrapolated by zero-order extrapolation over a predetermined macrostep width, the mass receives a time-delayed excitation. The resulting dynamic reaction of mass is therefore affected, which can lead to stability problems. In this case, the smaller macrostep width results in less time delay and thus, in some cases, stabilizes the integrated system response. The choice of macrostep width and extrapolation method is therefore important for high quality, high value physical integrated system response.

The couplers of the first and second partial systems are constructed based on the coupled network, the first partial system information and the second partial system information, the execution sequence, the extrapolation method and the macro step width, and the joint simulation is a macro. Performed during the time step.

A partial system has a partial model that mimics a real model (eg, the part itself or a part flow model). The model describes the response of the subset through algebraic and / or differential relationships. This partial model is created and simulated using a simulation tool (eg, a CAD program). The integrated system is assembled from multiple subsystems so that the integrated system can be modeled and simulated, and thus the truth value representation of the integrated system's response in the real world. Each system solves a given system area (flow model, structural model, temperature distribution) of the integrated system. The individual subsystems influence each other. For example, a predetermined temperature distribution reveals a flow model or structural model (eg, various deformation responses of the structural model) that depends on it.

Input parameters are parameters that the solution algorithm requires as inputs to determine the simulation results or output parameters. The input parameters are, for example, temperature, geometric data, force, rotation speed, peripheral parameters (eg, outside air), flow, etc. required by the solution algorithm.

The solution algorithm (solution) performs the desired simulation in the subsystem. The first solution algorithm or the second solution algorithm may then be the same or different. Moreover, the individual solution algorithms for the subsystems can use various fixed or variable step widths to solve the individual subsystems. The solution algorithm represents a numerical method, which can be used to determine the input parameters and the output parameters from the modeling subsystem.

The output parameters of the subsystem are predetermined values that are calculated and simulated using the relevant solution algorithm. For example, if a dynamic transformation response modeled by a differential equation must be simulated in a partial system, for example this may be a geometric value for geometric model formation. For example, the stress U of the algebraic partial model U = R * I can be simulated, in which the resistance R is an internal model parameter in the partial system and I is the current and thus the input parameter. The input parameter I can be calculated in another subsystem, which is the corresponding output parameter at that location.

The coupled network determines which subsystems must be coupled to each other and how. The coupled network forms a coupler between the two subsystems and determines which of the first and second parameter outputs is defined as the amount of coupling for the corresponding first and second parameter inputs. To do. The parameter output of a subsystem may be assigned to multiple parameter inputs of another subsystem.

Couplers can be implemented wired or wirelessly to exchange information.

The subsystem information is the information of the subsystem, and the information characterizes the subsystem. Partial system information can optionally be controlled and matched, or is affected by coupling with other subsystems. The partial system information is, for example, direct access, I / O occasional operation, instantaneous frequency and / or simulation time, as detailed below.

Subsystem information such as "direct access" or "input / output ad hoc operation" may be pre-available or well-known to individual subsystems. For example, this information, within the scope of additional information, can be provided by a subsystem creator (eg, a simulation tool) along with the subsystem. In this case, the analysis cost for analyzing the partial system for each individual subsystem is therefore reduced.

Moreover, an execution sequence is selected that determines in which sequence the first parameter output and the second parameter output are defined with each other. In other words, it determines which subsystems are running when (ie, simulated over a default macro time step). For example, if the subsystems influence each other, first run a given subsystem and then assume the same macro time step for subsequent subsystems with the resulting output parameters without subtracting the input parameters. , It is advantageous to use (sequential execution). The execution sequence can be given in advance by the user and / or can be matched in the configuration process.

In addition, the extrapolation method establishes the first and second parameter inputs during the macro step width (eg, during the coupling time point). Various extrapolation methods can be used for the individual first and second input parameters. These values must be supplemented (estimated) and interpolated accordingly to calculate the output parameters, especially if the parameter inputs are not freely available for the actual macro time steps. For example, since various extrapolation methods are suitable for solving a partial system for a predetermined solution algorithm, proper selection of the extrapolation method has a positive effect on the simulation result of the partial system.

The couplers of the first and second partial systems are constructed based on the coupled network, the first partial system information and the second partial system information, the execution sequence, the extrapolation method and the macro step width. By considering the above factors in the coupling of two or more subsystems, the configuration of the subsystems is based on the confirmed information with respect to each other (eg, execution sequence, coupled network, etc.). The configuration can be adjusted or improved on its own (eg, macrostep width, selection, extrapolation method), thus improving the simulation results of the integrated simulation. In addition, particularly matched and improved coupling can reduce the duration and resource consumption of integrated simulations.

Users of collaborative simulation today face the challenge of constructing the resulting integrated simulation. Within the configuration, the user should define, for example, a subsystem execution sequence and preliminarily provide extrapolation methods and associated macrostep widths for a large number of coupled signals (eg, input or output parameters). .. The configuration is improved by using the method according to the invention, especially by considering the partial system information, the coupled network, the execution sequence, the extrapolation method and the macro step width in the configuration.

According to another exemplary embodiment of the method, the collaborative simulation is terminated after the macro time step, or the collaborative simulation is newly performed. When this method is newly interactively executed or co-simulated, the part detected as a result of the simulation in order to match the execution sequence, extrapolation method and macro step width when performing the next simulation. System information (eg, direct access, I / O occasional operation, instantaneous frequency, simulation time, etc.) and coupled networks and configurations may be used. Therefore, the improved configuration of the coupling of the partial systems is executed for each implementation, and the result of the joint simulation is optimized for each implementation.

According to other excitation examples, after the macro time step, an analysis of the coupling results (eg, discrete events, advanced system dynamics) is performed at the coupling time within the scope of the partial system analysis, and the analysis is performed. Based on this, the partial system information, execution sequence, externalization method and / or macro step width of the first and / or second partial system are matched or configured. For example, the steep curves of the coupling signal during the simulation in the subsystem between the coupling time points used improper execution sequences, improper extrapolation methods (thus improper input parameters were defined). , Or it means that the macro step width selection was inappropriate (eg, selected too large). When a new joint simulation is performed based on this information, the selection of the execution sequence and / or the selection of the extrapolation method and / or the selection of the macro step width can be matched.

Within the analysis of the partial model, the input and output parameters (coupling signals) of the partial system are analyzed for various characteristics during the execution time. The input and output parameters of the partial model are brought out for analysis of the partial model, for example in individual parts. In addition, so-called "direct access" determination, "input / output ad hoc operation", subsystem simulation time and "frequency analysis" (instantaneous frequency), and subsystem calculation time and discrete events can be considered.

According to another exemplary embodiment, the subsystem information is between the first input parameter and the first output parameter of the first subsystem and the second input parameter and the second of the second subsystem. Has input / output at any time with the output parameters of.

The dynamic characteristics of the subsystem between all available input / output combinations of the individual subsystems are considered as the input / output ad hoc behavior of the subsystem. For example, a subsystem may have one input / output ad hoc operation between inputs and outputs, or an additional input and output of the subsystem or an input / output ad hoc operation different from that associated with the additional subsystem. .. These input / output ad hoc operations of the subsystem can be performed by a MIMO (multiple inputs and multiple outputs) system or a method for system recognition of an input / output specific SISO (single input single output) based on the data. .. The occasional operation of this subsystem substantially determines what kind of solution algorithm or step width must be used to obtain a stable and accurate numerical solution. In the method according to the invention, the input / output ad hoc operation of the subsystem is substantially used to select the execution sequence and the individual extrapolation method.

According to an additional exemplary embodiment, the subsystem information has a simulation time of the first subsystem and / or the second subsystem. For example, the first subsystem has a different simulation time than the second subsystem (eg, by using different macro step widths). Basically, depending on the individual simulation time of the subsystem, for example, the macro step width or the execution sequence can be matched to improve the implementation of the joint simulation.

Supplementally, the situation is reserved that the described methods and devices for constructing the joint simulation can also be used within the scope of the real-time preliminary joint simulation. Unlike (non-real-time) collaborative simulations, real-time collaborative simulations allow partial systems to be simulated in real-time, so they are additionally used according to real-time (eg, wall clocks) at predetermined and combined time points. Need to be present for (eg, supplementary).

According to an additional exemplary embodiment, each macro in the subsystem has the computational time required to perform a simulation of the individual subsystem for each macro time step. Time response matching is performed to run the joint simulation in real time, referring to the computational time required by the subsystem at each join point using time steps. Recognize the lack of time response to run a joint simulation in real time by referring to the computational time required by the subsystem at each join time using each macro time step of the subsystem. This can be done, and the shortage can be stopped by constructing the joint simulation advantageously during the execution time. For example, in the configuration process, it may be considered that, for example, a partial system requires a shorter calculation time (as partial system information) than a macro time step (eg, defined in real time). In this case, in order to improve the collaborative simulation with the configuration, for example, the execution sequences of the subsystems may be matched, or flexible computational capacity may be incorporated in other ways.

According to additional exemplary embodiments, the partial system information has the instantaneous frequencies of the first and / or second input parameters and / or the first and / or second output parameters.

During the "frequency analysis" of the coupled signals, the individual coupled signals are tested for their frequency. The inspection is narrowed down to the detection of so-called instantaneous frequencies, and the instantaneous frequencies can be detected by a method derived from signal processing. For example, the instantaneous frequency is determined by the so-called Hilbert-Fang conversion.

According to additional exemplary embodiments, the subsystem information has direct access to the first and / or second input parameters of the subsystem.

The concept "direct access" represents a system characteristic in control technology, and fluctuations in system inputs result in fluctuations (actions) in system outputs directly and without delay. Detection of "direct access" can be done, for example, by a data-based system identification method. In connection with collaborative simulation, the incorporation of subsystems using "direct access" results in algebraic loops, and thus difficult or higher-level solution processes during collaborative simulation. For this reason, the recognition of the existence of direct access between the input and output volumes of a subsystem is of great interest and is therefore used for methods of automatically configuring collaborative simulations.

According to an additional exemplary embodiment, the subsystem information is a combination event (discrete event, higher order) of the first and / or second input parameters of the subsystem and / or the first and / or second output parameters. Has an analysis of the system's occasional operation).

According to an additional exemplary embodiment, the step of determining the macro step width is to determine the first macro step width of the first subsystem and the second macro step width of the second subsystem. Have. The first macro step width gives in advance a first coupling point in time, respectively, where the first output parameter can be determined at that time, and the second macro step width, respectively, where the second output parameter is at that time. A second binding time point, which can be determined by, is given in advance.

Collaborative simulations include various subsystems that can be simulated over different macrostep widths in stages. By utilizing different macro step widths, different simulation times for individual subsystems are also generated.

The simulation time describes how long the individual subsystems have been simulated and can be obtained directly from, for example, time stamped coupled signal data (ie, output parameters).

According to an additional exemplary embodiment, the step of determining the extrapolation method can be determined using that method during the first parameter input during the first macrostep width (during the time of coupling). , The determination of the first extrapolation method of the first partial system and the second parameter input can be determined using that method during the second macro step width (during the time of coupling). It has the determination of the second extrapolation method of the second partial system. Various extrapolation methods can be used for the individual first and second input parameters.

According to an additional aspect of the present invention, an apparatus is described that constitutes a joint simulation for an integrated system having at least one first subsystem and a second subsystem. The first partial system has at least one first parameter input and at least one first parameter output, and based on the first parameter input, the first parameter is used using the first solution algorithm. The output is determinable. The second subsystem has at least one second parameter input and at least one second parameter output, and based on the second parameter input, the second parameter is used using the second solution algorithm. The output is determinable.

The apparatus has a coupling unit for defining a coupling network, and the network couples a first subsystem and a second subsystem with a coupler, and either of the first and second parameter outputs is Determine if it is defined as the amount of coupling for the corresponding first and second parameter inputs.

The apparatus further provides the first subsystem information (direct access, input / output ad hoc operation, instantaneous frequency, simulation time, subsystem calculation time and discrete events) of the first subsystem and the second subsystem of the second subsystem. It has a confirmation unit for confirming partial system information. The apparatus further has a selection unit for selecting an execution sequence that determines in which sequence the first parameter output and the second parameter output are mutually defined. The apparatus further comprises an extrapolation unit for determining the extrapolation method, in which the first and second parameter inputs can be determined using the method during the macrostep width (during the time of coupling). .. The apparatus further has a step width unit for determining a macro step width (single) or a macro step width (plural), and the step width unit is between a first subsystem and a second subsystem. Preliminarily gives a coupling time point at which the exchange of the corresponding first and second input parameters and the first and second output parameters in.

The apparatus further constitutes the coupling of the first and second subsystems based on the coupling network, the first and second partial system information, the execution sequence, the extrapolation method and the macro step width. It has a component for performing joint simulations of the first and second subsystems.

Various units, particularly the coupling unit, the selection unit, the extrapolation unit, the step width unit, the component, and the like of the present apparatus may be manufactured as one processor. It is also possible to configure any combination of these or other units or a large number of units as a common processor. All units may be made as one common processor.

According to an additional exemplary embodiment, a computer-readable storage medium is described in which a program for constructing a collaborative simulation for an integrated system is stored therein, and the program is executed by a processor. Perform or control the method described above.

According to additional exemplary embodiments, program elements for constructing a collaborative simulation for an integrated system are described, which, when executed by a processor, perform or control the methods described above.

In summary, the present invention describes a method of automatically constructing a collaborative simulation. The automatic configuration of the joint simulation takes into account the quality of the simulation results and, in summary, uses the following findings: The simulation tool execution sequence determines which coupling signals must be extrapolated; Extrapolation increases as the macrostep width increases during non-repetitive joint simulations; extrapolation methods are well suited for coupling in a variety of applications.

Abstractly and generally speaking, in collaborative simulation, inaccessible subsystems are integrated into an integrated system by combining subsystem inputs and subsystem outputs. Therefore, the information needed to configure a collaborative simulation is generally not accessible to the user. The coupled network is the information available and describes which subsystem inputs are coupled to which subsystem outputs. In combination with the commonly available coupled network, this method analyzes the relevant subsystems during run time, thereby extracting additional information for automatic configuration of collaborative simulations.

The method according to the invention additionally assembles detailed information on the extrapolation method used. Specially, here we refer to the mathematical model of the extrapolation method that enables the derivation of the so-called "effective band".

The method according to the invention, which automatically configures the joint simulation, is free to consider additional useful information. For example, here information about numerical solutions algorithms that are subordinate and used to simulate individual subsystems is quoted. Extrapolation of the input amount of the partial system can have undesired effects on the underlying solution algorithm, such as loss of execution time or numerical problems. For example, information on the occasional operation of the subsystem can be considered.

Collaborative simulations generally consist of at least two interacting subsystems. It is important to solve the causality problem by extrapolating the amount of coupling by exchanging only and therefore limited data at discrete time points (joining time points) during joint simulation. In this case, the execution sequence essentially determines which coupling amount (eg, input parameter) must be compensated for, and at what position the coupling error is introduced into the integrated system. For example, this is discussed below assuming the same macro time step width for the relevant subsystems. If all subsystems are calculated in parallel, all coupled signals in the combiner must be extrapolated over a defined macro time step. On the contrary, if the first subsystem is calculated before the second subsystem and then the third subsystem, then only the coupling signal in the second coupler is with the third subsystem. It must be extrapolated in each joint simulation process with the second subsystem.

The coupling signal is extrapolated to solve the causality problem within the scope of the joint simulation. Various extrapolation methods, comprehensive measures to correct coupling errors, use coupled signal data available at the actual time. The choice of extrapolation method for a particular application can be left to the user of the collaborative simulation. The method according to the present invention constructs a mathematical model of an individual extrapolation method and uses the extrapolation features derived from it. An example of such a feature is the so-called "effective band" which can be determined from the transfer function of the individual coupler. The macro time step width defines the discrete time points (join time points) during the joint simulation where the data exchange between the subsystems takes place at that time point. Since different subsystems can have different occasional movements and different extrapolation methods are used, basically different choices of macro time step widths are significant.

The embodiments of the present invention are in computer programs, i.e. using software and using one or more specific electric switches, i.e. in hardware (eg, FPGA or ASIC), or in any hybrid form, i.e. software. It can be realized using parts and hardware parts. Subsystems can be simulated, for example, on a computer locally (also divided into various computer cores) or in a network that is topologically divided into various computers.

It is pointed out that the examples described herein are merely a limited selection of possible different embodiments of the invention. A number of different embodiments have been explicitly disclosed to those skilled in the art using explicit different embodiments, as the characteristics of the individual embodiments can be combined with each other in an appropriate manner. Can be regarded as a thing. In particular, some embodiments of the present invention are described in the device claims, and other embodiments of the present invention are described in the method claims. However, it will be immediately apparent to those skilled in the art when teaching the present application, and unless otherwise specified, in addition to the combination of features belonging to the type (singular) of the subject (singular) of the present invention, the subject (plural) of the present invention. Any combination of features belonging to various types (plurality) of is also possible.

Examples are described in detail below with reference to the accompanying drawings for additional explanation and for a better understanding of the present invention. Shown below:
A schematic diagram of a joint simulation for an integrated system according to an exemplary embodiment of the invention is shown. A schematic diagram of the flow of the framework according to the exemplary notation of the method according to the invention is shown. A schematic diagram of an exemplary embodiment of the present invention is shown. A schematic diagram of extrapolation between two binding time points is shown.

The same or similar parts in the various figures have the same reference numbers. The notation in the figure is schematic.

FIG. 1 shows a schematic diagram of a joint simulation for integrated system 100 according to an exemplary embodiment of the invention. A joint simulation is assembled from the first partial model 110, the second partial model 120 and the third partial model 130. The first partial system 110 has at least one first parameter input 111 and at least one first parameter output 112, and the first parameter output 112 is the first based on the first parameter input 111. It can be determined using the solution algorithm 114 of.

The second partial system 120 has at least one second parameter input 121 and at least one second parameter output 122, and the second parameter output 122 is second based on the second parameter input 121. Can be determined using the solution algorithm 124 of.

The third partial system 130 has at least one third parameter input 131 and at least one third parameter output 132, and the second parameter output 132 is a third based on the third parameter input 131. It can be determined using the solution algorithm 134 of.

Each of the partial systems 110, 120, 130 has a partial model that mimics a real model (eg, the component itself or a flow model of the component). The model describes the responses of the partial systems 110, 120, 130 via algebraic and / or differential relationships. This partial model is created and simulated using simulation tools 113, 123, 133 (eg, CAD program). The integrated system 100 is assembled from a plurality of subsystems 110, 120, 130 so that the integrated model 100 can be modeled and simulated, and thus the truth value representation of the response of the integrated system 100 can be made in the real world. Each system 110, 120, 130 solves a predetermined system area (flow model, structural model, temperature distribution) of the integrated system 100. The separate subsystems 110, 120, 130 influence each other. For example, a predetermined temperature distribution reveals a flow model or structural model (eg, various deformation responses of the structural model) that depends on it.

Subsystems 110, 120, 130 can be simulated in a network that is locally (divided into various computer cores) computers or topologically divided into various computers.

Input parameters 111, 121, 131 are parameters that the solution algorithms 114, 124, 134 require as inputs to determine the simulation results or output parameters 112, 122, 132. Input parameters 111, 121, 131 are, for example, temperature, geometric data, strength, force, rotation speed, peripheral parameters (eg, external temperature), flow, etc. required by the solution algorithm.

The solution algorithms (solutions) 114, 124, 134 perform the desired simulation within the subsystems 110, 120, 130. The first solution algorithm 114 or the second solution algorithm 124 may then be the same or different. Moreover, the individual solution algorithms for the subsystems can use various fixed or variable step widths to solve the individual subsystems. The solution algorithms 114, 124, 134 represent a numerical method, which can be used to determine the output parameters 112, 122, 132 from the input parameters 111, 121, 131 and the modeled subsystems 110, 120, 130. ..

The output parameters 112, 122, 132 in the partial systems 110, 120, 130 are predetermined values calculated and simulated using the solution algorithms 114, 124, 134. Multiple values of output parameters 112, 122, 132 can also be determined during the macro time step.

A first coupler 101 is formed between the first partial system 110 and the second partial system 120. A first parameter output or output parameter 112 is obtained by the first subsystem 110 at a predetermined coupling time point and is given to the second subsystem 120 as a second parameter input or input parameter 121. In the second coupler 102, for example, a second output parameter of the second subsystem 120 is given as a third input parameter 131 in the third subsystem 130.

Further, the partial system may also have a plurality of input parameters 121 obtained from different submodels 110, 130, for example. In this example, for example, a third output parameter 132 in the third coupler 103 is given to the second subsystem 120 as a second input parameter 121. At the same time, a first output parameter 112 is given to the partial system 120 as an additional second input parameter via the first coupler 101.

An improved configuration of couplers 101, 102, 103 according to the method according to the invention may improve the joint simulation of the integrated system 100.

An exemplary flow of the method according to the present invention is shown in combination with the joint simulation derived from FIG. After the start 200, the partial system information 201 is first defined. In that case, the first partial system information (for example, direct access, input / output occasional operation, instantaneous frequency, simulation time) of the first partial system 110 and the second partial system information of the second partial system 120 are defined. These partial system information 201 are taken out from the databank or by the user's preset in the initial process. When this method is repeatedly executed later, the partial system information specified when this method was executed before can be taken out.

Further, the first subsystem 110 and the second subsystem 120 (or a number of additional subsystems) are coupled by the couplers 101, 102, 103 and the first and second parameter outputs 112, 122. A coupling network 202 is defined to determine which of the two is defined as the coupling amount for the corresponding first and second parameter inputs 111, 121.

Next, select the execution sequence 203, which determines in which sequence the first parameter output 112 and the second parameter output 122 are defined to each other, and thus which first and second to solve the causality problem. Determines if the parameter inputs 121, 131 must be extrapolated.

Next, an extrapolation method 204 is defined in which the first and second parameter inputs 112, 122 can be determined using that method during the macro step width (during the time of coupling).

Further, the macro step width 205 is the exchange of the corresponding first and second input parameters 111 and 121 between the first and second subsystems 110 and 120 and the first and second output parameters 112 and 122. Preliminarily gives a join time, at which the exchange will take place.

Finally, the couplers 101, 102, 103 of the first and second partial systems 110, 120 are the coupled network (202), the first partial system information and the second partial system information, the execution sequence (203), and the supplement. Constructed on the basis of extrapolation (204) and macro step width (205), the collaborative simulation is performed over macro time steps.

After the start 200 of the joint simulation, freely available information (eg, partial system information 201) is evaluated. For the first freely available information during the collaborative simulation, use default values, for example, or do not consider default values in this configuration. By superimposing freely usable information, for example, an execution sequence 203 can be determined. Next, based on this and freely available information, the appropriate macro step width 205 is selected in the final process to determine the appropriate extrapolation method 204 in the additional process, so that the macro time step determined by the joint simulation is followed. The configuration 206 of the joint simulation is fixed to the above. After this simulation 207, that is, at the time of the next coupling, the partial system analysis 208 is performed and the subsystem information 201 accumulated so far is updated. If the end of the joint simulation has not been reached (t <t _end ) after the simulation step, this process having each simulation step is repeated until the end of the joint simulation (t = t _end ) is reached.

FIG. 3 describes a possible technical switchover of how to automatically configure a joint simulation during (t <t _end ). The next simulation step 207 is performed after the configuration 206, in which the two subsystems 110, 120 are coupled via the coupler 101 for joint simulation, is actually completed, and therefore the macro time step is calculated. A new partial system analysis 208 is performed. In this step, the partial systems 110, 120, 130 are analyzed based on the combined data (input parameters 111, 121, 131 and output parameters 112, 122, 132), for example, direct access 302, input / output occasional operation 304, instantaneous frequency 305 and / Or relevant information such as simulation time 303 is extracted or finalized for this configuration. A method of selecting and extrapolating the execution sequence 203 from this database, along with available additional information such as, for example, the coupled network 202 and the individual solution algorithms 301 (114, 124, 134) of the corresponding subsystems 110, 120, 130. The selection of 204 and the selection of the macro time step width 205 are made. In additional step 206, these adjustments are used to construct a joint simulation (206). Then, for example, a macro time step is calculated (207) and the process repeats until the end of the joint simulation (t = t _end ).

In addition, FIG. 3 describes possible switchings for storing information, which serve as a database for automatically configuring collaborative simulations. The data available to automatically configure the collaborative simulation is stored, for example, in various matrices (eg, 202, 301, 302, 303, 304, 305). For example, the various matrices can be an I / O combined or combined network 202 of all related partial systems 110, 120, 130, available "direct access" 302 of the partial systems 110, 120, 130, partial systems 110, 120, 130. Input / output occasional operation 304, instantaneous frequency 305 of the coupling signal (eg, in couplers 101, 102, 103), solution algorithm 301 underneath if available, and / or subsystems 110, 120, 130. The actual simulation time 303 is described. This information is extracted during the run time and / or following the simulation.

In the matrix, for example, the columns form the parameter inputs of the subsystems 110, 120, 130, and the rows form the parameter outputs of the subsystems 110, 120, 130.

The proposed method analyzes local information (eg, partial system analysis), analyzes global information (eg, coupled network), and uses this information to construct a joint simulation as a whole.

FIG. 4 shows extrapolation between two binding time points. In a non-repetitive collaborative simulation, the relevant subsystems 110, 120, 130 are solved once exactly over each predetermined macro time step. The choices of execution sequence 203, extrapolation method 204 and macro time step width 205 must be determined at the time of join before calculation. If, for example, discrete events or higher-order system ad hoc behaviors occur during this macro time step, the joint simulation for this step was not configured in response to the system response. This situation is shown graphically in FIG. FIG. 4 shows the coupling signal 401, which is defined by the macro time steps of the solution algorithms 114, 124, 134 of the partial systems 110, 120, 130.

At coupling point 402, the coupling signal 401 is extrapolated by the primary extrapolation 404 until the next coupling time 403 over the macro time step to be calculated. The last two values from the history of the coupling signal 401 prior to the coupling time 402 are used. While calculating the macro time step, only the event 405 occurs at time _{t e} at subsystems 110, 120, which leads to strong fluctuations 401 at the time of the event 405, thus coupling a signal 401 relative to the extrapolated curve 404 It will deviate greatly. It is conceivable to repeat the simulation over this macro time step, but it is not really useful, because popular simulation tools, for example, often offer the possibility to reconfigure a partial system simulation on a previous join point. Because there isn't. Nevertheless, in order to ensure a suitable configuration, an extension of the automatic configuration according to the method according to the invention is used.

The automatic configuration of collaborative simulations may be limited in terms of effectiveness due to its non-repetitive nature. To this end, we propose an iterative joint simulation of the same joint simulation as an extension of this method, using the findings from the preceding simulation curve and (automatically) according to the execution sequence 203, extrapolation method 204 and The selection of the macro time width 205 is configured. Each position that needs to be acted upon in this configuration is already known from the history of partial system analysis 208 and is addressed for automatic configuration for subsequent joint simulations corresponding to the straight lines described above. In FIG. 2, this extension is reserved by an additional inquiry 210. If the condition "#-condition" is met in query 210, the configuration is automatic again after the joint simulation (t = t _end ) based on the available coupling signals and the partial system information generated from the partial system analysis 208. Condition "#-Condition" when the user activates this function, AND, the maximum number of iterations is reached, OR, and the predetermined quality of the joint simulation is reached (logical AND before logical OR). Is satisfied.

Supplementally, it is suggested that "contains" does not exclude other elements or processes, and "one (female)" or "one (male / neuter)" does not exclude many. Furthermore, it is suggested that the features or steps described with reference to one of the above examples can be used in combination with the other features or steps of the other examples described above. The reference numbers in the claims should not be considered limited.
[Item 1]
In a method of constructing a joint simulation for an integrated system having at least one first subsystem and a second subsystem.
The first partial system has at least one first parameter input and at least one first parameter output, and the first parameter output is the first solution based on the first parameter input. Can be determined using an algorithm
The second subsystem has at least one second parameter input and at least one second parameter output, and the second parameter output is the second solution based on the second parameter input. A method that can be determined using an algorithm
The above method
The first subsystem and the second subsystem are coupled by a coupler, and which of the first and second parameter outputs is the amount of coupling with respect to the corresponding first and second parameter inputs. To determine the combined network to determine if it is defined as
To determine the first partial system information of the first partial system and the second partial system information of the second partial system,
Select the execution sequence that determines in which sequence the first parameter output and the second parameter output are defined with each other, and which of the first parameter input and the second parameter input is extrapolated. Determining what must be done and
Determining the extrapolation method, in which the first and second parameter inputs can be determined using the method during the macro step width.
A combination in which the corresponding exchange of the first and second input parameters and the exchange of the first and second output parameters between the first subsystem and the second subsystem are performed at that time. Determining the macro step width that gives the time point in advance,
Based on the coupled network, the first and second partial system information, the execution sequence, and the extrapolation method and the macro step width, the coupler of the first and second partial systems. A method that includes configuring and running a collaborative simulation during macro time steps.
[Item 2]
The method according to item 1, wherein the joint simulation is terminated or the joint simulation is newly executed after the macro time step.
[Item 3]
After the macro time step, a partial system analysis is performed at the point of coupling, based on the first and second partial system information and the second partial system information of the first and second partial systems, as well as the combined network. The method according to item 1 or 2, wherein at least one of the execution sequence, the supplementary method, or the macro step width is matched.
[Item 4]
The first subsystem information and the second subsystem information are placed between the first input parameter and the first output parameter of the first subsystem, and the second subsystem of the second subsystem. Between the input parameter and the second output parameter, and between the third input parameter and the third output parameter of the third subsystem, the first subsystem, the second subsystem and the above. The method according to any one of items 1 to 3, which has an input / output operation of the third partial system at any time.
[Item 5]
Any one of items 1 to 4, wherein the first subsystem information and the second subsystem information have a simulation time of at least one of the first subsystem or the second subsystem. The method described in the section.
[Item 6]
The first partial system information and the second partial system information are the instantaneous frequency of at least one of the first input parameter or the second input parameter, and the first output parameter or the second. The method according to any one of items 1 to 5, which has at least one of the instantaneous frequencies of the output parameters of.
[Item 7]
The first partial system information and the second partial system information are the first output parameter or the second output of the first partial system, the second partial system and the third partial system. The method according to any one of items 1 to 6, wherein the first and second input parameters have direct access to at least one of the parameters.
[Item 8]
Since the first sub-system information and the second sub-system information have at least one calculation time of the first sub-system or the second sub-system, the first sub-system, the first sub-system, the above. The calculation required by the first subsystem, the second subsystem and the third subsystem at each connection time using the macro time steps of the second and third subsystems. The method according to any one of items 1 to 7, wherein the time response matching is performed in order to perform the joint simulation in real time with reference to time.
[Item 9]
The first subsystem information and the second subsystem information are the first input parameter of the first subsystem, the second subsystem and the third subsystem, or the second input parameter. The item 1 to 8, which comprises an analysis of at least one of the binding events and an analysis of at least one of the first output parameter or the second output parameter. Method.
[Item 10]
The process of determining the macro step width is
Determining the first macro step width of the first subsystem,
The first macro step width preliminarily gives a first coupling time point at which the first output parameter can be determined at that time.
Determining the second macro step width of the second subsystem above
The item according to any one of items 1 to 9, wherein the second macro step width has a second output parameter that can be determined at that time, and a second binding time that is given in advance. Method.
[Item 11]
The process of determining the extrapolation method is
Determining the first extrapolation method of the first subsystem, wherein the first parameter input can be determined using the method during the first macro step width.
Item 10 comprising determining a second extrapolation method of the second partial system, wherein the second parameter input can be determined using the method during the second macro step width. The method described in.
[Item 12]
A device that constitutes a joint simulation for an integrated system having at least one first subsystem and a second subsystem.
The first subsystem has at least one first parameter input and at least one first parameter output.
Based on the first parameter input, the first parameter output can be determined using the first solution algorithm.
The second subsystem has at least one second parameter input and at least one second parameter output, and the second parameter output is the second solution based on the second parameter input. Can be determined using an algorithm
The above device
The first subsystem and the second subsystem are coupled by a coupler, and which of the first and second parameter outputs is used as the coupling amount for the corresponding first and second parameter inputs. The coupling unit for determining the coupling network to determine whether it is defined, and
The determination unit of the first partial system information of the first partial system and the second partial system information of the second partial system, and
A selection unit for selecting an execution sequence, which determines in which sequence the first parameter output and the second parameter output are defined with each other.
An extrapolation unit for determining the extrapolation method, wherein the first and second parameter inputs can be determined using the method during the macro step width.
The exchange of the corresponding first and second parameter inputs and the exchange of the first and second parameter outputs between the first and second subsystems is performed at that time, in advance of the coupling time. Given, the step width unit for determining the macro step width,
Based on the combined network, the first partial system information and the second partial system information, the execution sequence, the extrapolation method, and the macro step width, the coupling of the first and second partial systems is configured. A device having a component for performing the joint simulation during the macro step width.
[Item 13]
The program according to any one of items 1 to 11 when a program for constructing a joint simulation for an integrated system is a computer-readable storage medium stored therein and the program is executed by a processor. A computer-readable storage medium that implements or controls the above method.
[Item 14]
A program for causing a processor to execute the method according to any one of items 1 to 11.

100 Integrated system 110 1st partial system 111 1st input parameter 112 1st output parameter 113 1st simulation tool (tool 1)
114 First solution algorithm (Solution 1)
120 Second partial system 121 Second input parameter 122 Second output parameter 123 Second simulation tool (tool 2)
124 Second solution algorithm (Solution 2)
130 Third partial system 131 Third input parameter 132 Third output parameter 133 Third simulation tool (tool 3)
134 Third solution algorithm (Solution 3)
101 1st coupler 102 2nd coupler 103 3rd coupler 200 Start 201 Confirm partial system information 202 Confirm combined network 203 Select execution sequence 204 Determine externalizing method 205 Macrostep width 206 Configure the joint simulation 207 Latest simulation process 208 Analysis 209 Decide on a new implementation 210 End decision 211 End 301 Determine the solution algorithm 302 Confirm direct access 303 Determine the simulation time 304 Input and output at any time Confirm the operation 305 Confirm the instantaneous frequency 401 Coupled signal 402 Coupled time 403 Coupled time 404 Primary supplement 405 Event

Claims (10)

In a method of constructing a joint simulation for an integrated system having at least one first subsystem and a second subsystem.
The first partial system has at least one first parameter input and at least one first parameter output, and the first parameter output is the first solution based on the first parameter input. Can be determined using an algorithm
The second partial system has at least one second parameter input and at least one second parameter output, and the second parameter output is the second solution based on the second parameter input. A method that can be determined using an algorithm
Causes combined in coupler and said second system part the method and the first system part, which of the first parameter output and the second parameter output, the corresponding first parameter input and was to determine the connection network to determine whether the defined as the amount of binding to the second parameter input, wherein the coupling network comprises a first system part and second system part, they It is determined by capturing the physical relationship between the parameter input and the parameter output of the above and the defined coupling amount of the parameter input and the parameter output .
The determination of the first partial system information of the first partial system and the second partial system information of the second partial system is to determine the first partial system information and the second partial system. The information is referenced from the database for configuration in the initial steps, and the first subsystem information and the second subsystem information are the simulation times of the first subsystem and / or the second subsystem. Including, and
Select an execution sequence that determines in which sequence the first parameter output and the second parameter output are defined with each other, and which of the first parameter input and the second parameter input is extrapolated. Determining what must be done , said that the execution sequence can be described by the operator and / or adapted in said steps of configuration .
Said first parameter input and said second parameter input can be estimated by using the method during a macro step width, the method comprising determining the extrapolation method,
Replacement of the corresponding replacement of the first parameter input and a second parameter input and the first parameter output and the second parameter output between the first system part and the second part system To determine the macro step width that will be executed at that point and give the join point in advance,
The coupling network, the first system part information and the second partial system information, said execution sequence, and on the basis of the extrapolation method and the macro-step width, wherein the first system part and second system part A method comprising constructing the coupler of the above and performing a joint simulation during a macro time step. Wherein the first system part information and the second partial system information, between the first parameter input and the first parameter output from the first system part, first of the second system part between two parameters input and the second parameter output, and the third third system part of the parameters input and the third parameter said first system part between the output, the first The method of claim 1 , comprising the input / output dynamics of the second subsystem and the third subsystem. Wherein the first system part information and the second partial system information includes at least one of the instantaneous frequency of said first parameter input or said second parameter input, the first parameter output or wherein at least one of the second instantaneous frequency of the parameter output method according to claim 1 or 2. Wherein the first system part information and the second partial system information, the first system part, the second system part and the third part system, the first parameter output or the second At least one said first direct reach through parameter input and said second parameter input possess to, the direct reach-through, said first system part and the second part of the parameters output Claims 1 to 3 refer to a change in the first parameter input and / or the second parameter input of the system that directly and without delay causes a change in the first parameter output and / or the second parameter output of the system. The method according to any one of the above. Since the first subsystem information and the second subsystem information have at least one calculation time of the first subsystem or the second subsystem, the first subsystem, said. The calculation required by the first subsystem, the second subsystem and the third subsystem at each coupling time using the macro time steps of the second and third subsystems. The method according to any one of claims 1 to 4 , wherein time-response matching is performed in order to perform the joint simulation in real time with reference to time. The step of determining the macro step width is
Determining the first macro step width of the first subsystem and
The first macro step width, the first parameter output are each determinable at the time, providing a first coupling point in advance,
Determining the second macro step width of the second subsystem and
Said second macro step width, the second parameter output are each determinable at the time, providing a second coupling point in advance, having, in any one of claims 1 to 5 The method described. The step of determining the extrapolation method is
Determining the first extrapolation method of the first partial system, wherein the first parameter input can be estimated using the method during the first macro step width.
A claim comprising establishing a second extrapolation method of the second partial system, wherein the second parameter input can be estimated using the method during the second macro step width. The method according to 6 . A device that constitutes a joint simulation for an integrated system having at least one first subsystem and a second subsystem.
The first subsystem has at least one first parameter input and at least one first parameter output.
Based on the first parameter input, the first parameter output can be determined using the first solution algorithm.
The second partial system has at least one second parameter input and at least one second parameter output, and the second parameter output is the second solution based on the second parameter input. Can be determined using an algorithm
The device
Wherein the first system part and the second part system is combined in combiner, which of the first parameter output and the second parameter output, the corresponding first parameter input and the first It is a coupling unit for determining a coupling network that determines whether or not it is defined as a coupling amount with respect to the parameter input of 2, and the coupling network is the first partial system, the second partial system, and theirs. A coupling unit determined by capturing the physical relationship between the parameter input and the parameter output and the defined coupling amount of the parameter input and the parameter output .
It is a determination unit of the first partial system information of the first partial system and the second partial system information of the second partial system, and the first partial system information and the second partial system information are stored in a database. Referenced from for configuration in the initial step, said first subsystem information and said second subsystem information includes simulation time for said first subsystem and / or said second subsystem. Confirmed unit and
A selection unit for selecting an execution sequence that determines in which sequence the first parameter output and the second parameter output are defined with each other, the execution sequence being described by the operator and /. Or with a selection unit that can be adapted in the above steps of configuration ,
It said first parameter input and said second parameter entered during the macro step width can be estimated by using the method, and extrapolation unit for determining an extrapolation method,
The corresponding first parameter input and exchange of the second parameter input and the first parameter output and exchange of the second parameter output thereof between the first system part and second system part A step width unit for determining the macro step width, which is executed at the time point, gives the join time point in advance, and
The coupling network, the first system part information and the second partial system information, said execution sequence, the extrapolation method, binding of the based on the macro-step width first system part and second system part A device having a component for performing the joint simulation during the macro step width. The computer-readable storage medium in which a program for constructing a joint simulation for an integrated system is stored therein, according to any one of claims 1 to 7 , when the program is executed by a processor. A computer-readable storage medium that implements or controls the above method. A program for causing a processor to execute the method according to any one of claims 1 to 7 .